How to read current going into an inverter

[Mickel]

I want to measure current being used by a 12V inverter. I have bought a
200 amp shunt and a labjack usb daq box. The problem I am having is that
the measured voltage across the shunt keeps jumping all over the place.
The reason being that the current to the inverter peaks 50 times per
second and the labjack is taking the measurements at different random
points on the sine wave. I can take, say, 250 measurements per second
and average them but that seems kindof overkill and will introduce
timing problems when windows starts doing unnecessary crap. I tried
putting a cap across the shunt but that had little effect. Is there
something I can do to smooth out that voltage? I notice the volt meter
gives a nice smooth reading that is exactly what I would expect.

Thanks in advance,
Michael

 

[Phil Allison]

"Mickel"

I want to measure current being used by a 12V inverter.

** You want to know the *average * current - right ??

I have bought a 200 amp shunt and a labjack usb daq box. The problem I am
having is that the measured voltage across the shunt keeps jumping all
over the place. The reason being that the current to the inverter peaks 50
times per second and the labjack is taking the measurements at different
random points on the sine wave. I can take, say, 250 measurements per
second and average them but that seems kindof overkill and will introduce
timing problems when windows starts doing unnecessary crap. I tried
putting a cap across the shunt but that had little effect. Is there
something I can do to smooth out that voltage? I notice the volt meter
gives a nice smooth reading that is exactly what I would expect.

** A simple RC network will average the current wave - like all DMMs have.

A 10 kohm resistor in series with the signal and a 1uF cap across the output
should smooth it nicely.

I hope your A to D has a 1M ohm input.


..... Phil

 

[Mickel]

"Mickel"

** You want to know the *average * current - right ??

Yep. I certainly don't need to know the current 250 times per second.

** A simple RC network will average the current wave - like all DMMs have.

A 10 kohm resistor in series with the signal and a 1uF cap across the output
should smooth it nicely.

Sounds easy enough, I will give it a try tomorrow.

I hope your A to D has a 1M ohm input.

I think it has something like that.

Thanks Phil for the quick reply

 

[Mickel]

A 10 kohm resistor in series with the signal and a 1uF cap across the output
should smooth it nicely.

Hi Phil, I might be doing something wrong but unfortunately that didn't
work. I presume I'm meant to put the resistor on the positive wire of
the signal (in series) and the cap across the pins of the daq (ie I
don't put the cap on the shunt side of the resistor). I'm using an 8k
resistor because that is all I could find and a 1uF electrolytic cap.
The cap is around the correct way. I've included an ascii representation
of my circuit and the data I'm getting back for 0.4 of a second. A
voltage of 0.4 on the daq represents zero because it's using an opamp on
the input. This is a part that was supplied with the daq box. The thing
I find interesting is the current peaks 50 times a second. I thought it
should peak 100 times a second?

Thanks again,
Michael

DAQ
- +
| |
+-||-+
| |
| \
| / 8k
| \
| |
shunt



0.4807
0.4454
0.4293
0.3958
0.4466
0.4795
0.4496
0.4233
0.3946
0.4454
0.4795
0.4556
0.4215
0.3958
0.4454
0.4783
0.4448
0.4245
0.3952
0.4448
0.4807
0.4514
0.4227
0.3964
0.4442
0.4789
0.4502
0.4215
0.3946
0.4442
0.4789
0.452
0.4227
0.3952
0.4442
0.4801
0.4502
0.4239
0.3946
0.4442
0.4795
0.455
0.4287
0.3946
0.4442
0.4789
0.4484
0.4209
0.3946
0.443
0.4795
0.452
0.4263
0.394
0.443
0.4795
0.449
0.4233
0.3928
0.443
0.4795
0.4502
0.4221
0.3928
0.443
0.4801
0.4472
0.4263
0.3934
0.443
0.4801
0.452
0.4251
0.3928
0.4424
0.4789
0.4544
0.4281
0.3934
0.4424
0.4795
0.452
0.4197
0.3934
0.4424
0.4789
0.4472
0.4275
0.3928
0.4424
0.4795
0.449
0.4287
0.3916
0.4418
0.4795
0.452
0.4215
0.3916
0.4418

 

[Phil Allison]

"Mickel"

Hi Phil, I might be doing something wrong but unfortunately that didn't
work. I presume I'm meant to put the resistor on the positive wire of the
signal (in series) and the cap across the pins of the daq (ie I don't put
the cap on the shunt side of the resistor). I'm using an 8k resistor
because that is all I could find and a 1uF electrolytic cap. The cap is
around the correct way. I've included an ascii representation of my
circuit and the data I'm getting back for 0.4 of a second. A voltage of
0.4 on the daq represents zero because it's using an opamp on the input.
This is a part that was supplied with the daq box. The thing I find
interesting is the current peaks 50 times a second. I thought it should
peak 100 times a second?

** Think I see the problem now.

You are trying to do a measurement with an inappropriate tool.

I suggest you remove the RC network and go back to direct sampling.

Then do the AVERAGING in software - a simple average is what is required
over say 100 samples.



...... Phil

 

[Mickel]

You are trying to do a measurement with an inappropriate tool.

I suggest you remove the RC network and go back to direct sampling.

I actually think your solution could work, it just needs some fine
tuning. The problem I think is that the voltage is very low so the cap
is not having much chance to do anything. I tried going to a 10uF 25V
cap and this made a big difference. It still wasn't enough but it was a
big change. The other option is that I could put the RC circuit on the
output of the op-amp where the voltages are 11x the input or more
depending on how it is set.

The cap I was originally using was a 1uF 63V electrolytic. Does the 63V
make much difference compared to, say, a 1uF 10V cap? Would the lower
voltage rating make any difference if the capacitance is the same? I
would have thought not you never know.

Then do the AVERAGING in software - a simple average is what is required
over say 100 samples.

I would really rather avoid this because it makes it very timing
dependent and I can just see the reliability dropping. If windows starts
doing updates at 4 sample/min then this will be very different to 250
per second. The device itself also seems to be less reliable in high
speed mode, I get a lot more errors etc if things are not perfect and I
have to worry about buffer sizes and other crap.

BTW, it looks like your advice from a previous thread about how to
measure 240V with this unit is good. The bit about making sure I put a
small load on the transformer. I found that without this load the unit
was showing 240V for a full 10 minutes after the 240V was switched off.

 

[Phil Allison]

"Mickel"

I actually think your solution could work, it just needs some fine tuning.

** Use as large a value cap as you like - and NO the voltage rating has no
effect.

But remember, a large cap will be slow to respond to changes.


..... Phil